Detecting the occurrence of suicide clusters at city block scale: evidence from a 26-year data series

Suicide clusters have profound negative impacts in the affected communities. Despite the fact that the occurrence of this phenomenon is a rare event, it has been observed across diverse cultural contexts. Suicide clusters are commonly described as a higher number of suicides than expected by chance, occurring at relatively small time/space scales. In particular, point clusters have been reported as occurring in areas ranging from hundreds of meters to several kilometers and lasting from a few days to several months or years. Nevertheless, effective prevention of suicide clusters requires robust estimation of their spatial and temporal scales, especially at short distances such as the city-block scale. This study analyzes data from the city of Valdivia, Chile (1996-2021) using a DBSCAN-based approach. We detected seven suicide clusters at very small scale in different years. In the clusters, the distances between suicides were <300 m (two to three city blocks). For the entire period, 6 % of the suicides occurred in a cluster. These clusters contain between three and four suicides, each with higher prevalence of men and people over 30 years old. Our results provide important insights for implementing preventive actions at the neighborhood scale.